# Use context for organization (optional but good practice)
context("Calculation Functions")

# --- Test calculate_daily_load ---
test_that("calculate_daily_load aggregates correctly", {
  # Create sample activity data
  sample_acts <- dplyr::tibble(
    id = c("1", "2", "3"),
    type = rep("Run", 3),
    start_date_local = lubridate::ymd_hms("2023-01-01 10:00:00", "2023-01-01 12:00:00", "2023-01-02 09:00:00"),
    date = lubridate::as_date(c("2023-01-01", "2023-01-01", "2023-01-02")),
    moving_time = c(1800, 3600, 2700), # 30, 60, 45 mins
    distance = c(5000, 10000, 8000), # 5k, 10k, 8k
    duration_mins = c(30, 60, 45),
    distance_km = c(5, 10, 8)
  )

  # Test aggregation by duration
  daily_duration <- calculate_daily_load(sample_acts, load_metric = "duration_mins")
  expect_equal(nrow(daily_duration), 2) # Should have 2 unique dates
  expect_equal(daily_duration$load[daily_duration$date == "2023-01-01"], 90) # 30 + 60
  expect_equal(daily_duration$load[daily_duration$date == "2023-01-02"], 45)

  # Test aggregation by distance
  daily_distance <- calculate_daily_load(sample_acts, load_metric = "distance_km")
  expect_equal(daily_distance$load[daily_distance$date == "2023-01-01"], 15) # 5 + 10
  expect_equal(daily_distance$load[daily_distance$date == "2023-01-02"], 8)

  # Test handling of empty input
  empty_result <- calculate_daily_load(sample_acts[0, ], load_metric = "duration_mins")
  expect_equal(nrow(empty_result), 0)
  expect_true("date" %in% names(empty_result))
  expect_true("load" %in% names(empty_result))

  # Test that missing dates are filled with 0
   sample_acts_gap <- dplyr::tibble(
     date = lubridate::as_date(c("2023-01-01", "2023-01-03")),
     duration_mins = c(30, 60)
   )
   daily_load_gap <- calculate_daily_load(sample_acts_gap, load_metric = "duration_mins")
   expect_equal(nrow(daily_load_gap), 3) # Jan 1, 2, 3
   expect_equal(daily_load_gap$load[daily_load_gap$date == "2023-01-02"], 0)

})

# --- Test calculate_acwr ---
test_that("calculate_acwr computes rolling means and ratio", {
  # Create sample daily load data (enough for chronic period)
  set.seed(123)
  days <- 40
  sample_load <- dplyr::tibble(
    date = lubridate::today() - lubridate::days(days:1),
    load = sample(50:150, days, replace = TRUE)
  )

  acwr_result <- calculate_acwr(sample_load, acute_period = 7, chronic_period = 28)

  # Check output structure
  expect_true(all(c("date", "acute_load", "chronic_load", "acwr") %in% names(acwr_result)))
  expect_equal(nrow(acwr_result), days - 28 + 1) # Should lose first chronic_period-1 rows

  # Check a specific calculation (manual calculation needed for verification)
  # For day 28 (first valid row):
  # acute = mean(load[22:28])
  # chronic = mean(load[1:28])
  manual_acute_day28 <- mean(sample_load$load[22:28])
  manual_chronic_day28 <- mean(sample_load$load[1:28])
  manual_acwr_day28 <- manual_acute_day28 / manual_chronic_day28

  expect_equal(acwr_result$acute_load[1], manual_acute_day28)
  expect_equal(acwr_result$chronic_load[1], manual_chronic_day28)
  expect_equal(acwr_result$acwr[1], manual_acwr_day28)

  # Test handling insufficient data
   short_load <- sample_load[1:20,] # Less than chronic period
   expect_warning(calculate_acwr(short_load, chronic_period=28))
   empty_acwr <- suppressWarnings(calculate_acwr(short_load, chronic_period=28))
   expect_equal(nrow(empty_acwr), 0)

})

# --- Test Placeholder Functions (Example: TSS) ---
test_that("calculate_tss requires user_ftp and calculates placeholder", {
    sample_acts <- dplyr::tibble(
        id = "1", type="Ride", date=lubridate::today(),
        moving_time = 3600, duration_hours = 1, average_watts = 200
    )
    # Expect error if ftp is missing
    expect_error(calculate_tss(sample_acts, user_ftp = NULL), "requires 'user_ftp'")

    # Expect warning about placeholder
    expect_warning(tss_result <- calculate_tss(sample_acts, user_ftp = 250), "placeholder")
    expect_true("tss_load" %in% names(tss_result))
    # Check placeholder calc: 1hr * (200/250)^2 * 100 = 1 * 0.8^2 * 100 = 64
    expect_equal(tss_result$tss_load, 64)
})

# --- Test categorize_load_exposure ---
test_that("categorize_load_exposure assigns zones correctly", {
    sample_acwr <- dplyr::tibble(
        date = lubridate::today() + 0:4,
        acwr = c(0.5, 1.0, 1.4, 1.6, NA) # Under, Sweet, Caution, Danger, NA
    )
    categorized <- categorize_load_exposure(sample_acwr)
    expect_equal(categorized$acwr_zone, c("Underload", "Sweet Spot", "Caution Zone", "Danger Zone", NA))
})

# ADD MORE TESTS for other calculation functions (EF, Decoupling, PBs)
# These will be more complex and might require sample stream data.